Risk identification is useful in several contexts. As an example, service data can improve the balance between device security and data access. This task relies on computing accurate risk scores for acquired service data. The increasing number of new types of data and the increasing accessibility to different data sources provide additional challenges to risk identification. A common technical issue associated with risk identification is that time streams of service data are requested and received through multiple calls between systems. For example, a number of calls can be proportional to an amount of service data, such that an increase in the amount of service data prolongs processing time associated with a risk identification, and otherwise inefficiently consumes computing resources.